Diagnostic Studies in the Liver Transplant Candidate
Laboratory tests obtained in the immediate preoperative period should include measures of hepatic synthetic and excretory function, and should specifically include serum albumin and bilirubin. Routine coagulation studies (PT, PTT, fibrinogen, and platelet count) give a measure of how hepatic synthetic failure has affected clotting factor levels. The platelet count can be low in the case of portal hypertension, as platelets are sequestered in the enlarged spleen.
Preoperative measurement of serum electrolytes is essential to provide baseline data about possible hyponatremia, hypo- or hyperkalemia, hypocalcemia, and hypomagnesemia. BUN and creatinine concentrations should be measured to assess the patient's renal function, which may in turn presage poor platelet function if uremia is present. This assessment of renal function (and acid–base status; see below) guides the possible use of intraoperative renal replacement therapy. Although rarely required, intraoperative renal replacement therapy may be needed if significant renal failure, electrolyte abnormality, or acid–base problems are found in the immediate preoperative laboratory investigation.
Hemoglobin should be measured to establish the preoperative baseline oxygen-carrying capacity. The patient's blood type should be redetermined, and a sample sent to the blood bank for antibody screening in anticipation of allogeneic transfusion. An electrocardiogram should be obtained in the immediate preoperative period to search for new changes suggestive of coronary artery disease. Finally, a preoperative chest x-ray is obtained as a baseline for comparison of postoperative films.
Patients presenting for liver transplantation almost always have an extensive body of laboratory and diagnostic test results available to aid in perioperative risk assessment and anesthetic planning. Patients with acute liver failure may be exceptions to this generalization, being listed for transplantation and subsequently receiving grafts within a few days of their initial presentation.14 This may also be true of patients previously evaluated as transplant candidates but moved up on the transplant list due to an abrupt worsening of their MELD score. However, in almost all cases, the following assessments are either repeated upon admission for transplantation or have been performed within the prior year.
Laboratory assessment of pulmonary function includes arterial blood gas measurement and pulmonary function testing. Blood gas measurement identifies patients with hypoxia, suggesting hepatopulmonary syndrome. Pulmonary function testing often reveals small lung volumes reflecting the effect of ascites. In patients with a long history of smoking, there may be a superimposed obstructive pattern on spirometry. These results are important, as the liver transplant patient is mechanically ventilated for a prolonged intraoperative period and perhaps for some time in the postoperative period.
Cardiac assessment of the patient being considered for liver transplantation focuses on functional and invasive tests of cardiac performance that assess ischemic potential and the search for cardiac structural anomalies that might compromise outcome from orthotopic liver transplantation.150
The incidence and severity of coronary artery disease increases with age and with well-known risk factors, many of which are commonly found in patients with liver failure such as diabetes, smoking, and obesity. Furthermore, better surgical and anesthetic techniques are allowing older patients—that is, patients with greater age-related risk of significant coronary disease—to be considered for liver transplantation. Flow-limiting coronary artery disease is of grave concern because liver transplantation is still a highly stressful operation. The potential for sudden, massive, and ongoing blood loss complicated by extreme electrolyte derangements is still present, so all potential liver transplant recipients must be evaluated as if they will be required to tolerate minutes to hours of a hypothetical state with a heart rate above 110 beats/min and a mean arterial pressure of less than 50 mm Hg, with a hemoglobin of 7 g/dL and a pH under 7.2. Only patients with ideal cardiac status are likely to survive this scenario unscathed.
To ensure that new and previously listed transplant candidates meet these criteria, the liver transplant anesthesia team periodically reviews the preoperative evaluation of individuals high on the liver transplant list. This task involves reviewing the echocardiogram, the results of functional stress testing, and, if performed, cardiac catheterization results. This is not an academic exercise, as a significant fraction of patients being evaluated for liver transplantation have coronary artery disease, and the coronary disease of many of these patients was unsuspected prior to pretransplantation evaluation. For example, in a study applying coronary angiography to all potential liver transplant recipients older than 50 years of age, there was a 27% incidence of moderate to severe coronary artery disease, with 13.3% of the cohort having clinically unsuspected moderate or severe coronary disease.151 Another study showed an incidence of 5.6% for coronary artery disease in patients older than 40 years of age,152 and the consensus seems to be that about 2.5% to 10% of patients have moderate to significant coronary artery disease.153 Untreated significant coronary disease is potentially lethal in the setting of liver transplantation.150 In 1 study, half of all patients with coronary artery disease who underwent liver transplantation died in the perioperative period, and the morbidity rate was 81%.154
It appears that dobutamine stress echocardiography or possibly exercise stress thallium imaging is the best screening test in patients with end-stage liver disease.153,155-157 A recent meta-analysis indicates that dobutamine stress echocardiography has a superior negative predictive value in patients having elective noncardiac surgery.158 Dobutamine stress is commonly used in pretransplant evaluation because it is considered to most closely mimic the state commonly found during liver transplantation and in end-stage liver disease.155 A negative dobutamine stress echocardiogram with adequate stress appears to predict a favorable perioperative cardiac outcome.156,159 It is important to achieve at least 85% of the predicted maximal heart rate so that the dobutamine stress echocardiogram is diagnostic. Otherwise the diagnostic value of the test is compromised.160 However, a recent retrospective analysis of OLT patients who underwent both dobutamine stress echocardiography and coronary angiography prior to their transplants showed that in patients who did achieve target heart rates, dobutamine stress echocardiography had a low sensitivity (13%) and a low positive predictive value (22%), questioning the accuracy of stress echocardiography in this population and the possible need for an alternative method of cardiac risk stratification.161
In the hyperdynamic state produced by the dobutamine stress echocardiography protocol, a significant number of patients develop a dynamic left ventricular outflow tract obstruction. However, despite the fact that more patients with dynamic left ventricular outflow tract obstruction developed intraoperative hypotension at transplantation, the overall outcomes (mortality) were the same between patients with obstruction and those without.162
Most transplant programs use a cardiac risk stratification schema similar to that devised by Plevak155; the version used by our program, for example, is shown in Fig. 57-8.
Suggested algorithm for assessment of inducible myocardial ischemia. OLT, orthotopic liver transplantation.
The ideal management of potential liver transplant recipients with significant coronary disease is a difficult problem with little evidence to guide decision making. Patients with end-stage liver disease who undergo coronary artery bypass grafting have significant morbidity. For example, in 1 study, patients with cirrhosis who underwent coronary artery bypass grafting had a 58% incidence of morbidity and significant complications.163 Even patients with relatively mild liver disease (Child class A and B liver disease) who underwent coronary artery bypass grafting had extremely high mortality and morbidity.164 The benefit of an intervention, such as coronary angioplasty, stenting, or atherectomy, has not been formally studied on a large scale to show outcome compared with coronary artery bypass grafting.153 Small series have reported good results with combined orthotopic liver transplantation and coronary artery bypass grafting,152,165,166 but at present, such a combined procedure is considered heroic in most transplant centers.
Independent of ischemic coronary artery disease, many patients with end-stage liver disease have a poorly understood disorder known as cirrhotic cardiomyopathy (a different entity from alcoholic cardiomyopathy) despite supranormal cardiac outputs.30 This condition is multifactorial, and its mechanisms may include impairment of the β-adrenergic system, nitric oxide (overproduced in liver failure), cytokines, and the prolonged hyperdynamic circulation.167 Many patients with apparently normal ventricular function prior to surgery develop left ventricular failure in the postoperative period.33,159 These patients, representing 1% to 6% of liver-transplanted individuals, may have had occult cirrhotic cardiomyopathy. Unfortunately, prospective diagnostic criteria for cirrhotic cardiomyopathy are lacking, although a recommendation made at the 2005 World Congress of Gastroenterology in Montreal suggested that a classification system of cirrhotic cardiomyopathy includes systolic and/or diastolic dysfunction in response to stress in the absence of any preexisting cardiac disease.168
All patients being considered for liver transplantation should undergo a screening transthoracic echocardiogram. This is a good screening test to search for cardiac structural abnormalities as well as anomalies of the surrounding vasculature. It is also an important test to exclude portopulmonary hypertension.
Cardiac structural anomalies can affect outcome by impairing cardiac performance intraoperatively (eg, functional or fixed valve stenoses) or by allowing passage of emboli from the right to the left side of the heart. Patent foramen ovale (PFO) and other septal defects may be a significant risk to orthotopic liver transplant patients and patients undergoing hepatic resection when fixed or transient right-to-left shunting occurs. Right-to-left shunt increases the possibility of paradoxical embolus of clot, air, or debris. Hepatic surgery can lead to large openings in the inferior vena cava, so the embolization problem can be severe. In fact, in 2 series, between 1% and 6% of all patients undergoing orthotopic liver transplantation had thrombus detected in the right heart.169,170 Because roughly 25% of all hearts in unselected autopsy subjects have PFOs171 the risk of injury due to paradoxical embolism may be greater than currently appreciated. On the other hand, in a recent study evaluating the risk of hepatic transplant patients with patent foramen ovales, 27 patients with patent PFOs were compared to 61 non-PFO transplant patients, and both perioperative outcomes and 30-day mortality rates were similar in both groups.172 This does not prove that PFO patients are not at elevated risk. Their anesthesia and surgical teams might simply have avoided accidental emboli better. Therefore, in preparation for elective hepatic surgery and certainly in the evaluation for orthotopic liver transplantation, cardiac septal defects should be sought by echocardiography or magnetic resonance imaging (MRI).
Transthoracic echocardiography can be used to interrogate the intra-atrial septum using color-flow Doppler mode. Additionally, bubble contrast echocardiography is a highly sensitive test for septal defects. Early passage (within 1-2 beats) of air from the right to the left heart indicates a septal defect, whereas later arrival of air on the left side indicates intrapulmonary shunting. This latter finding cannot be remedied, but when atrial defects are detected, consideration should be given to closing them before transplantation.
Closure of patent foramen ovales prior to orthotopic liver transplantation is controversial, although it can now be accomplished percutaneously.173 No controlled studies evaluating the usefulness of this intervention have appeared. Factors favoring closure are (1) embolism is common, particularly during the reperfusion period; (2) right heart dysfunction with right heart pressures greater than left heart pressures is common during reperfusion; and (3) paradoxical embolism occurs in this setting.174 On the other hand, the relationship between the presence or size of a patent foramen ovale and the likelihood of paradoxical embolus has not been clearly established. Although there are case series of paradoxical embolism documented by transesophageal echocardiography during liver transplantation,174 it is not clear whether these necessarily occur through a patent foramen ovale. Furthermore, closing a patent foramen ovale commits the coagulopathic patient with liver disease to months of antiplatelet agents.
Magnetic resonance imaging has recently been applied to the assessment of the intra-atrial septum. Small studies comparing this technique with color-flow Doppler echocardiography and bubble contrast studies indicate that MRI has similar sensitivity and specificity to these echocardiographic techniques.175
The final cardiopulmonary problem to be investigated in the prelisting evaluation is portopulmonary hypertension, the management of which has been described previously. Figure 57-1 outlines the diagnostic and therapeutic decision approach to portopulmonary hypertension. Portopulmonary hypertension occurs in 2% to 4% of patients with end-stage liver disease and 5% to 10% of patients being evaluated for orthotopic liver transplantation.53 The clinical predictors of pulmonary hypertension in liver transplantation include systemic hypertension, right ventricular dilation by echocardiography, estimated pulmonary artery systolic pressure of 40 mm Hg or greater by transthoracic echocardiography, right ventricular hypertrophy by echocardiography, or a right ventricular heave.176 The mortality associated with pulmonary hypertension during liver transplantation is significant. Half of the patients with mean pulmonary pressures between 35 and 50 mm Hg died in the peritransplant period, and mortality was 100% in those with mean pulmonary artery pressure greater than 50 mm Hg.177 These results are similar in a multicenter retrospective review,58 leading many centers to deny orthotopic liver transplantation to patients with more than mild portopulmonary hypertension. However, in some centers, mild portopulmonary hypertension (ie, mean pulmonary artery pressure <35 mm Hg with good ventricular function) is not associated with poor early or late outcomes.178 A retrospective study examining outcomes in portopulmonary hypertensive patients demonstrated 5-year survivals of 14%, 45%, and 67% for patients in 3 subgroups, respectively, who had (1) no therapy or liver transplantation alone, (2) only therapy for portopulmonary hypertension, or (3) therapy for portopulmonary hypertension followed by liver transplantation.179
Transesophageal echocardiography is rarely indicated in the preoperative assessment of orthotopic liver transplantation candidates. The procedure may require sedation, which can be complicated in the patient with end-stage liver disease. Furthermore, there is the risk of disrupting esophageal varices during the examination.
Operative Approaches for Liver Transplantation
Liver transplantation is conveniently divided into 3 phases: the preanhepatic phase, the anhepatic phase, and the neohepatic or reperfusion phase. During the preanhepatic phase, a complete hepatectomy is performed. During the anhepatic phase, vascular anastomoses between the donor liver and the recipient's vessels are constructed. During the neohepatic phase, the hepatic arterial and biliary anastomoses are constructed, and the wound is closed.
A liver transplant begins like a hepatic resection, using the same incision and exposure and imposing the same hemodynamic consequences. The surgeon exposes and gains control of the hepatic vasculature. The hepatic artery, portal vein, and the common bile duct are clamped and divided. Two different techniques are commonly used for controlling hepatic venous outflow and constructing this anastomosis: (1) an en-bloc technique in which part of the recipient's vena cava is resected and replaced with a section of the donor vena cava, or (2) a "piggyback" technique in which the recipient vena cava is clamped with a side-biting clamp and a side-to-side anastomosis is constructed to the donor liver venous outflow. The choice of technique has important implications for anesthetic management and the expected course of the case, as described in the following.
In the en-block resection technique, the inferior vena cava must be clamped above and below the liver, with consequent severe reduction in venous return to the heart. Because both the portal vein and the inferior vena cava are clamped, the entire body below the caval cross-clamp, as well as the abdominal viscera, suffers venous congestion and ischemia unless an alternate route of venous return can be established. This massive loss of venous return, coupled with the insult of recirculating blood from such a large ischemic territory at the time of reperfusion, leads to profound hemodynamic instability requiring pharmacologic intervention, sudden hypothermia at reperfusion, and elevated potential for malignant cardiac arrhythmias. These difficulties have motivated the development and routine use of venovenous bypass techniques. This is commonly achieved via venovenous bypass from the lower half of the body to a great vein draining into the superior vena cava (Fig. 57-9).
Schematic diagram of a venovenous bypass circuit.
Venovenous bypass may be used to decompress the lower systemic venous circulation, the portal circulation, or both. The advantages and disadvantages of venovenous bypass have been reviewed.180 Without the preserved venous return afforded by venovenous bypass, the cardiac output and blood pressure may fall dramatically; often the cardiac output falls by more than 50%. Such falls in cardiac output are suggested to increase morbidity and mortality,180 although primary research has failed to detect this association.181,182
Failure to decompress the lower systemic and portal venous systems leads to severe venous congestion below the caval cross-clamp. Above the clamp, reduced venous return reduces left ventricular preload and cardiac output, with a net result of reduced perfusion pressure for organs below the caval cross-clamp. Thus, without venovenous bypass, renal venous pressure is elevated, systemic arterial pressure is depressed, and renal perfusion pressure is degraded. Thus venovenous bypass might be expected to afford some protection from renal failure in the perioperative period, but the net effect on renal function is equivocal. For example, venovenous bypass is not associated with improved postoperative renal function relative to matched patients whose liver transplant was performed without bypass.183
First described in 1984,184 venovenous bypass is a technique wherein blood is actively pumped using a centrifugal pump from large-bore drainage cannulae to a similarly large return line (Fig. 57-10). These can be inserted percutaneously or via a cut-down procedure. A common drainage site for the lower systemic venous circulation is the left femoral vein. The portal vein, if drained, is accessed from the surgical field. Common return sites are the right internal jugular and left subclavian veins via percutaneous access or the left axillary vein via a cut-down technique.
Large-bore venovenous bypass cannula with guidewire and dilators.
Percutaneous insertion of the bypass cannulae may be performed by the surgical team or by the anesthesia team at the beginning of the case. An alternative approach is to prep the left groin and axilla for possible cut-down access and then defer access until it becomes clear that venovenous bypass is needed. Percutaneous access appears to be quicker and to provide better flows,185,186 but the cut-down technique persists. The bypass cannulae are very large, and when placed percutaneously, require multiple passes with successively larger sharp, stiff dilators passed over a stiff guidewire to create a large enough skin entrance, as illustrated in Fig. 57-11.
Subclavian placement of a percutaneous venovenous bypass cannula. The left panel demonstrates the use of a large-bore dilator after placement of the guidewire into the central circulation. The center panel shows the cannula in place and clamped prior to connecting to the venovenous bypass pump tubing. The right panel shows a portable chest x-ray confirming intravascular placement.
Assiduous care should be taken to pass the dilators only deep enough to enlarge the skin punctures, and under no circumstance should the tip of any dilator be allowed to pass beyond the end of the guidewire. Exsanguinating hemorrhage into the chest and cardiac tamponade from cardiac puncture have occurred as mishaps of venovenous bypass cannula placement.187 Extreme care should be taken to facilitate reliable cannulation of the target vein, including the possible use of ultrasonography to guide initial venous access and manometry to confirm venous placement using a small, temporary catheter prior to beginning dilation. A postinsertion chest x-ray to confirm vascular placement and correct insertion depth is mandatory prior to initiating venovenous bypass.
The bypass circuit should be carefully cleared of air as connections are made, and all connections should be secured prior to initiating bypass. Systemic heparinization is not performed, as the bypass circuit is heparin bonded.188 Bypass flow rates of 1.5 to 5 L/min are typical. The initiation of bypass is a critical period, as malplacement of the return line becomes apparent with disastrous consequences almost immediately. For example, accidental placement of the tip of the return line through a cardiac chamber wall and into the pericardial sac will lead to near-instantaneous, severe cardiac tamponade with initiation of flow. Fatal air embolism has occurred due to connection failure during venovenous bypass, so the pump and its circuit require constant attention by dedicated, specialized personnel.
Despite the improved conditions afforded during surgery, it is not clear that venovenous bypass improves long-term outcomes. Because of the potential for complications, coupled with improvements in surgical techniques (see below) some have questioned the usefulness of routine venovenous bypass.180
The alternative approach to liver transplantation, the vena cava preservation (or "piggyback") technique,189 is designed to preserve vena cava flow for all but a few minutes of the transplant procedure. Portal venous and hepatic arterial control is obtained as above. Instead of a cross-clamp, a "side-biter" clamp placed on the inferior vena cava isolates the liver from the systemic venous system. Figures 57-6 and 57-12 illustrate this clamping method. In Fig. 57-12, a button of vena cava with the hepatic vein attachments has been removed, and the hole is held closed by the side-biter clamp at the bottom of the wound. The vena cava deep to this clamp is largely open to flow. This allows venous return from the lower systemic but not the portal venous system. Frequently, a brief period of total venous occlusion with a vena cava cross-clamp is needed to position the side-biter clamp. Vena cava distension due to overzealous volume replacement during the hepatectomy can make the piggyback technique more difficult for the surgeons. Because exposure for the hepatectomy is more difficult and the anastomosis is more complicated, the piggyback technique may take longer to perform. However, many programs now use this technique almost exclusively.190
Photograph of a side-biting cross-clamp (lower left) on the inferior vena cava after removal of the liver. The patient's head is to the upper left.
Piggyback transplantation is associated with a lower incidence of renal failure,191 better gas exchange, and better acid–base status at reperfusion,192 as well as better intraoperative hemodynamic stability. A retrospective, 3-year analysis of 426 liver transplantations utilizing either retrohepatic caval resection with venovenous bypass, piggyback technique with venovenous bypass, or piggyback technique without venovenous bypass demonstrated that the piggyback-only technique utilized the smallest amount of blood products, had the lowest incidence of renal failure, and had overall better patient and graft survival when compared to the other 2 techniques.193 Additionally, this technique results in little to no need for venovenous bypass, as demonstrated in a series of 500 liver transplantations in which the piggyback technique was utilized and no patients required venovenous bypass in order to maintain hemodynamic stability.194
Failure to decompress the portal system may lead to splanchnic congestion unless the recipient has large portosystemic shunts. Nevertheless, the piggyback technique reduces the transfusion requirement as well as the need for vasoactive drugs during the completion of the hepatectomy and construction of the venous anastomoses.195 Temporary portocaval shunts have been described in association with the piggyback technique. The temporary shunt appears to improve intraoperative hemodynamics and reduce transfusion requirements,196 but is not widely used. In general, use of the piggyback technique dramatically reduces the severity of problems during the anhepatic and reperfusion stages of the operation.
As the hepatectomy is being completed, a separate surgical team prepares the donor liver for implantation, working at a separate table. The graft vessels are trimmed to fit their counterparts in the recipient. Next, the graft is brought to the recipient surgical field, and the vascular anastomoses are constructed. The surgical goal during this phase of the operation is to construct patent, nonleaking anastomoses as efficiently as possible to minimize the warm ischemic time of the liver. Thrombosis of the recipient portal vein is a relatively common finding, either during the dissection or at the time of initial graft reperfusion. When discovered, portal venous thrombosis typically prompts an attempted thrombectomy using Fogarty catheters, a potentially bloody process. Typically, the donor hepatic to recipient vena cava and the donor portal to recipient portal venous anastomoses are constructed, flushed, and opened as quickly as possible to establish a circulation in the graft. Then the hepatic arterial anastomosis is constructed, flushed, and opened, completing the graft blood supply.
Finally, a biliary drainage is constructed. This may either be by direct anastomosis of the graft bile duct to the recipient common bile duct or via a Roux-en-Y construct using the jejunum.
Anesthesia for Liver Transplantation
Concerns and activities during anesthesia for liver transplantation mirror the major phases of the surgery. The anesthesia can thus be roughly divided into the following phases:
- Induction, preparation for surgery, and maintenance
- Preanhepatic phase
- Anhepatic phase
- Venous reperfusion of the graft
- Neohepatic phase
- Emergence/transport to ICU
The anesthesia team attends to multiple homeostatic goals throughout the case (Table 57-2). Liver transplant anesthetics are some of the most complex currently performed because of the surgical and medical criticality of the patients and because of the complexity of the equipment required to perform the case.
Table 57-2 Major Foci of Intraoperative Management during Anesthesia for Orthotopic Liver Transplantation ||Download (.pdf)
Table 57-2 Major Foci of Intraoperative Management during Anesthesia for Orthotopic Liver Transplantation
|Major Focus (Problem or Organ System)||Subtopics||Perturbing Factors|
Right ventricular performance
Hemorrhage (demand and carrying capacity)
Acid/base and electrolyte disturbances
Emboli, hepatic washout
Portal-to-central venous gradient
Graft perfusion is tenuous
Need to keep CVP adequate to provide cardiac preload during vena cava cross-clamping
Need to prevent venous air embolism
Hypovolemia and hypotension
Preexisting renal disease
Need for intraoperative renal replacement therapy
Large incision, large prepped area
Dilution due to massive hemorrhage
Absent gluconeogenesis during anhepatic phase
Hypocoagulable state due to liver failure
Platelet dysfunction due to uremia
Acute fibrinolysis at graft reperfusion
Elevated by K load in blood products
Hyponatremic patients at risk for central pontine myelinolysis with massive infusions of Na+-rich fluids
Citrate in blood products chelates Ca+2
Pulmonary vascular resistance
May be impaired by pleural effusions
May be impaired by intrapulmonary shunts
Acutely elevates in the face of acid load at graft reperfusion
Vasodilated hyperdynamic circulation
Vena cava compression
Possible venovenous bypass
Figure 57-13 is a panorama from the anesthesiologist's perspective during a liver transplant. Multiple user interfaces, cables, monitoring lines, and infusions must be well organized and systematically set up to make the cognitive load from the equipment acceptable and predictable so that the anesthesiologist can attend properly to the patient and the operation. Liver transplantation consumes tremendous resources. Most programs responding to a recent survey reported assigning at least 2 anesthesiology personnel and at least 2 additional professional personnel (ie, perfusionist, monitoring nurse, auto-transfusion nurse) to each case.197 Thus, in addition to the complexity of the equipment, team leadership, communication, and delegation of responsibility must be considered and managed during the case.
Panorama of anesthesia equipment for liver transplantation. The equipment (from left to right) includes rapid infusion system; 6-pressure physiologic monitor with 8-channel chart recorder; syringe microinfusion pumps for vasoactive drugs; forced-air warmer; syringe microinfusion pumps for anesthetics, muscle relaxants, etc; level-of-consciousness monitor; anesthesia machine; and delivery system for inhaled nitric oxide. Out of view to the right: blood recovery system.
During the preinduction phase, the final evaluation of the patient is performed. Although there is unlikely to be any doubt about the identity of the patient and the operation to be performed, the anesthesia team must independently confirm the patient's identity, the operation to be performed, the recipient's blood type, and any allergies. Last-minute laboratory results should be reviewed. It is always prudent for a member of the anesthesia team to personally consult with the blood bank about the upcoming liver transplant so that it may prepare for a possible heavy demand for blood products. This is particularly true if the patient has developed antibodies to allogeneic red blood cells from previous transfusions, presenting an added challenge to the blood bank.
Close communication with the graft harvest team is essential to best sequence the activities with the donor to minimize graft ischemic time while at the same time minimizing the recipient's exposure to risks of anesthesia and invasive monitoring. Thus one might obtain initial peripheral intravenous (IV) access and then wait to place any additional monitors or vascular access until confirmation that the graft is suitable. In most situations, graft suitability is known at least 1 hour before the need to start the recipient's surgery, leaving ample time for induction of anesthesia, placement of central venous or pulmonary artery catheters, positioning of the patient, application of warming devices, and the initial surgical prep and drape.
Most patients with end-stage liver disease have at least some encephalopathy. Thus it is prudent to dose sedative premedications, such as benzodiazepines and opioids, judiciously, giving small doses and monitoring closely for desired and undesired effects.
Monitoring for liver transplantation relies on standard monitors plus invasive monitors directed toward the likely areas of additional concern during the case. Central venous pressure monitoring is often needed to assess intravascular volume status, as well as to goal-direct low CVP technique when utilized. Additionally, central venous catheterization provides large-bore transfusion access, allows passage of a PA catheter if needed, and provides a conduit for central vasopressor administration. Thus placement of a central venous catheter for liver transplantation seems almost obligatory. However, central venous pressure monitoring gives much of the same information as provided by intraoperative transesophageal echocardiography (TEE), which could replace the intravascular monitor if another suitable route can be found for the drugs.
Pulmonary artery catheterization is carried out if left or right heart performance is in doubt, if problems with pulmonary hypertension are anticipated, and to distinguish hypotension due to hypovolemia or heart failure from that due to lack of peripheral vascular tone, especially if TEE is not utilized during the case. One type of pulmonary artery catheter also allows for A-V pacing. The pulmonary artery catheter allows one to follow cardiac output and stroke volume throughout the case. The initial cardiac output measurements typically confirm a hyperdynamic circulation, with cardiac outputs of 10 to 11 L/min frequently seen at rest in adult male patients with end-stage liver disease.
An arterial catheter is usually inserted to closely monitor blood pressure and to provide convenient access for frequent laboratory sampling. When compared with femoral arterial catheters, radial artery catheters may not reflect the true systolic pressure during the reperfusion phase of transplantation or during vasopressor administration.198 However, mean pressures measured in the radial and femoral arteries correlate well regardless of the phase of the case or vasopressor administration.198 Thus a radial arterial catheter is sufficient to guide the management of the mean pressure. In some centers, 2 arterial catheters are placed so that blood pressure monitoring may proceed without interruption during frequent arterial blood gas withdraws, or as backup in case of an arterial line failure.
Transesophageal echocardiography should be considered for assessment of cardiac performance during transplantation, either as an adjunct or in place of pulmonary artery catheterization, and for monitoring embolism of air or thrombus. Air embolism is common, and TEE frequently reveals such events. Transesophageal echocardiography is most useful for early detection of large embolic events, with the hope of minimizing their effect by virtue of early detection and removal of the source. It is also a useful tool to assess the effect both of embolic events and therapeutic interventions on cardiac performance.
It is also useful to consider level-of-consciousness monitoring (BIS™ or similar). Liver transplant patients may be very sensitive to anesthetics, requiring smaller doses than typical patients. Level-of-consciousness monitoring can thus be used to minimize exposure to anesthetic agents and their potential deleterious effects on the circulatory system while providing some assurance that the patient will not recall the surgery.
Vascular access for rapid volume replacement can either be peripheral or in the central circulation. The patient likely requires central venous access for vasopressor administration, so it is tempting to use this route for volume as well. However, it is important to isolate vasopressors from the main volume cannula, as fluid for resuscitation may be delivered under pressure, effectively stopping the vasopressors. Multilumen large-bore central venous catheters or multiple catheters at separate central venous sites provide the necessary isolation. Alternatively, peripheral vascular access (eg, multiple 14-gauge IV catheters or a rapid infusion catheter as described previously) may be established. It is important to have at least 1 dedicated cannula that can carry the maximum flow rate of the available rapid infusion device (see below).
Anesthetic Induction, Preparation for Surgery, and Maintenance
Because most patients with end-stage liver disease have ascites and/or have had sclerotherapy of lower esophageal varices, rapid-sequence induction is the norm, with cricoid pressure applied and with the patient in optimal position for laryngoscopy. Careful denitrogenation is important, as most end-stage liver disease patients have reduced functional residual capacity. In patients with poor cardiac function, a preinduction arterial line may be warranted. Vasodilation and encephalopathy make patients with end-stage liver disease sensitive to induction agents. Thus the induction dose of hypnotic drugs, such as propofol or thiopental sodium (Pentothal), should be reduced (eg, 1 mg/kg-1 of ideal body weight for propofol in a cachectic patient with encephalopathy). Some clinicians elect to sidestep this issue entirely by using an induction agent such as etomidate. Skeletal muscle paralysis is typically obtained with succinylcholine followed by a nondepolarizing agent, or with rocuronium at a dose sufficient for rapid-sequence induction.
Antibiotics targeted against skin flora (eg, cefazolin) should precede skin incision by no more than 60 minutes.199 Antibiotics should be redosed throughout the case to account for elimination and loss due to hemorrhage.
During maintenance of anesthesia, hepatic encephalopathy reduces opioid and anesthetic requirements. Thus it is possible to inadvertently administer relative overdoses of anesthetics by overestimating the patient's needs. Similarly, opioids may have enhanced potency and prolonged duration of action due to hepatic failure. On the other hand, massive hemorrhage and resuscitation can diminish the circulating opioid concentration. Thus it is appropriate to titrate intermediate acting opioids, such as morphine, hydromorphone, or fentanyl, to the patient's needs throughout the case. The choice of paralyzing agent should also be guided by the clinical situation. For example, agents requiring hepatic degradation are probably not optimal if early extubation is contemplated.
Positioning the patient requires special attention. Liver transplants are long cases with the potential for hypotension and consequent extremity hypoperfusion. Furthermore, surgeons may lean heavily on the patient to perform parts of the operation. Thus positioning injuries are possible, and great care should be taken to protect the patient by generously supporting and padding all of the body, either with compressible foam or viscoelastic gel. As surgical needs may warrant frequent changes in table position throughout the case, all extremities should be carefully secured. Pneumatic compression stockings are applied, and a urinary catheter is inserted.
The case is performed with the patient in supine position. If required, the right arm may be tucked alongside the patient's body to allow access for several surgeons. If so, the right arm is out of reach of the anesthesiologist, and no critical monitors or access devices should be placed there unless no alternative sites are available. In addition to problems with accessibility, devices placed in the tucked right arm may cause patient injury due to compression. Accordingly, we typically limit devices in the right arm to a single large-bore peripheral IV and arterial line. The left arm is abducted to facilitate access for venovenous bypass. Therefore, the left arm is also an ideal site for arterial monitoring and peripheral venous access. Venous access in the left arm may be occluded if venovenous bypass is instituted via cut-down to the axillary vein, but the bypass circuit has a separate high-flow inlet into which this infusion can be attached. Other sites of access include both femoral regions. Venous access below a potential vena cava cross-clamp is of little utility, either for monitoring or volume administration. Furthermore, the left femoral vein is typically reserved for possible venovenous bypass access. However, cannulation of the right femoral artery may be useful if a complicated case is anticipated. Both internal jugular veins are available with the patient in supine position. If required, dual large-bore catheters may be placed in a single internal jugular vein. Care must be exercised to ensure that the catheters are sufficiently separated along the axis of the vein to prevent forming a larger, single connected hole.
The entire abdomen and chest are prepped for surgery. Additionally, the left groin and axilla are included in 1 large, contiguous prep if venovenous bypass via open access is contemplated.
Patient warming becomes a major issue due to the extensive use of alcohol-based skin prep solutions over a large surface, the size of the incision, the duration of the case, and reperfusion of a cold graft. Hypothermia should be avoided because it negatively affects coagulation,121 resistance to infection,120 and the potential for early extubation. Warming strategies commonly focus on IV fluids and forced-air warming devices.
Forced-air warming is effective, although the extensive area exposed and prepped during liver transplantation makes finding sufficient surface area on which to apply warmers a challenge. A procedure involving the entire abdomen, as well as the left groin and axilla for establishment of venovenous bypass, leaves the head, left arm, and lower legs for warming. Forced-air warming blankets for underbody use, which presumably function by forming a warmed-air plenum under the drapes, are available and should be considered. These appear to be effective to the extent that they are not compressed and deflated by the weight of drapes and surgeons pressing against them. In this instance, we place forced-air warming blankets on all available sites, including under the patient. In our experience, patient temperature drops to about 36°C during induction, prepping, and draping; increases to 37°C during the preanhepatic phase; and then falls by approximately 1°C at hepatic reperfusion. Subsequently, the temperature increases to 37°C during the neohepatic phase, prompting the gradual discontinuation of forced-air warming to maintain normothermia. If lower-extremity forced-air warming is used, anesthesiologists must remember to cease use of these if aortic cross-clamping occurs at any time during the case, or thermal injury may result.
Warming intravenous infusions is mandatory, given the volume of refrigerated banked-blood products that will likely be transfused. Fluid warming for liver transplantation is usually achieved with commercially available high-flow warming devices. The ideal transfusion device has a reservoir to allow the user to establish a reserve of blood for sustained rapid infusion in the face of uncontrolled hemorrhage. The infusion flow rate should be reported. The device must contain an air detector that stops the infusion when air is detected in the patient limb of the circuit. Ideally the device contains a debris filter between the reservoir and the patient.
Three devices are in common use, and each has unique advantages and weaknesses. The characteristics of the 3 major rapid infusion systems are listed in Table 57-3. The first, manufactured by Level-1 (Rockland, MA, USA) uses a counter-current heat exchanger to warm fluid administered under pressure from bags. The second device, formerly manufactured by Haemonetics (Braintree, MA, USA), is no longer commercially available but is still in common use. The final device is manufactured by Belmont Instrument Corp. (Billerica, MA, USA) and is called the Belmont Fluid Management System (FMS). The pumped devices are pressure limited at 300 mm Hg; that is, their maximum flow rates in actual clinical use are limited by the resistance characteristics of the infusion catheter or by the maximum rate of the pump if the back-pressure is less than 300 mm Hg when maximum pump flow is attained. Of the 3 devices listed in Table 57-3, the RIS most closely approaches the characteristics of the ideal transfusion device. It has a debris filter/air removal system downstream of the warmer and pump but proximal to the final air detector, thus providing maximal embolism exclusion. The RIS also has the highest flow rate, delivering 1500 mL/min, or roughly one-third of the normal adult cardiac output, exceeding the flow capacities of all but the largest catheters. However, the RIS disposable insert is bifurcated, allowing the pumped flow to be sent to 2 infusion catheters in parallel, increasing maximum flow and volume delivered as compared to a single set due to a decrease in resistance.
Table 57-3 Characteristics of Common Rapid Infusion Devices ||Download (.pdf)
Table 57-3 Characteristics of Common Rapid Infusion Devices
|Device||Reservoir||Pump||Air Detector||Postpump Filter||Measured Flow Rate||Max Flow Rate|
|Level 1||No||No||Yes||No||No||Catheter limited|
|Haemonetics RIS||Yes||Yes||Yes||Yes||Yes (from pump)||1500 mL/min|
|Belmont FMS||Yes||Yes||Yes||No||Yes (from pump)||750 mL/min|
Of the 2 available devices, the FMS is preferable in that it satisfies more of the requirements listed above. Its maximum flow rate does not exceed the capacity of a RIC or a percutaneous introducer sheath, but it does exceed the capacity of a 14-gauge peripheral IV or the auxiliary lumen of a centrally inserted device such as the MAC (Multi-Access Catheter, Arrow International, Reading, PA, USA) or AVA (Advanced Vascular Access, Edwards Lifesciences, Irvine, CA, USA).
As in a hepatic resection, clamping the portal triad reduces venous return. In the hepatic resection, as mentioned above, cardiac output and blood pressure do not necessarily fall. Liver transplantation almost always requires a vena cava cross-clamp, creating total vascular exclusion physiology for at least a brief duration. In contrast to portal triad clamping, total vascular exclusion of the liver via a vena cava cross-clamp reduces systemic blood pressure and cardiac output because the loss of venous return is quite large. The vena cava cross-clamp may be applied briefly to facilitate completion of the hepatectomy or for a longer time to allow construction of the hepatic venous anastomoses. Severe cardiovascular compromise during the vena cava cross-clamp may necessitate institution of venovenous bypass. Prior to committing to vena caval and portal resection, the surgeon will often "test clamp" the cava and portal vein to ensure that the patient will tolerate hepatectomy without venovenous bypass. However, more moderate decreases in cardiac output and blood pressure can be treated by vasopressor administration coupled with judicious intravascular volume expansion. Modest doses of a vasopressor during caval cross-clamp to construct caval anastomoses are apparently well tolerated, with graft and patient survival rates comparable with those obtained when venovenous bypass is used.200
Low central venous pressure does not seem advantageous during liver transplantation, in contrast to hepatic resection. For patients undergoing orthotopic liver transplantation, keeping central venous pressure less than 5 mm Hg may be associated with higher incidences of postoperative renal failure and 30-day mortality relative to patients whose central venous pressure was allowed to run between 7 and 10 mm Hg, although more recent studies indicate that it may in fact be beneficial in that it decreases intraoperative blood loss, protects graft function, and has no detrimental effects on renal function in patients undergoing OLT.201,202
As in partial hepatectomy, 2 major contributors to hemodynamic changes during liver transplantation are changes in cardiac performance due to episodic alterations in venous return (discussed above) and hemorrhage. However, these factors are only 2 of the many sources of cardiovascular instability during liver transplantation. Sepsis, acidemia, hypocalcemia due to citrate toxicity, embolism, and acute right ventricular failure are all major contributions to hemodynamic problems. The period of graft reperfusion (described below) is particularly unstable in many instances.
The goals of hemodynamic management are to provide sufficient circulating volume, vascular resistance, and cardiac output to perfuse the vital organs. This is not guided by any single parameter but rather by a synthesis of all available data, including urine output, central venous pressure (in relation to the preoperative central venous pressure and with a full appreciation of the factors that may artificially alter it), and the absence of a vasopressor requirement except during periods of inadequate venous return.
With adequate volume management, vasopressors are rarely required intraoperatively except during periods of inadequate venous return (ie, during inferior vena cava cross-clamping) and transiently at liver reperfusion. Patients with higher MELD scores (>30) are more likely than those with lower MELD scores to require vasopressors.203 Administration of vasopressors is a choice between the global detrimental effects of untreated significant hypotension and the potential negative effect of vasoconstrictors on perfusion of the new graft. In animal models, both epinephrine and norepinephrine reduced graft macro- and microperfusion,204 but the functional consequences of these effects on human allograft performance or survival are unclear.
Vasopressin is a tempting choice because it is effective regardless of the pH, it reduces norepinephrine requirements, and it is effective for managing hepatorenal syndrome before transplantation. Most studies of vasopressin to date are in patients with septic or postcardiotomy shock. Vasopressin is effective in these circumstances, reducing the requirement for catecholamine vasopressor support. Limited observational studies indicate that although vasopressin raises the blood pressure, it does not compromise the microvascular circulation.205 In 1 study retrospectively reviewing the use of vasopressin in septic shock, vasopressin was associated with increased liver enzyme levels, bilirubin, or both.206 However, a recent study utilizing vasopressin in 16 patients undergoing liver transplantation showed that an infusion of vasopressin started after hepatic artery clamping and before caval clamping yielded a significant decrease in both portal vein pressure and blood flow but without a concomitant decrease in either cardiac output or intestinal perfusion.207
Terlipressin, a vasopressin analogue available in Europe and Asia, has been used in patients with cirrhosis and portal hypertension, demonstrating decreases in both hepatic and renal arterial resistance and portal venous blood flow, but without changes in portal vascular resistance.208
Adjuvant drugs, such as aprotinin, reduce the requirement for vasopressors, presumably by improving clotting performance and minimizing blood loss.209 However, these advantages must be balanced against the other potential risks of aprotinin infusion (see below).
Transfusion and Other Therapies to Maintain Intravascular Volume
The goal of volume management and transfusion in liver transplantation is to maintain sufficient intravascular volume to support a well-functioning circulation. Additionally, in many instances, the hemorrhage is sufficiently large that the anesthesiologist must intervene directly to control the composition of the circulating intravascular volume. Thus, in addition to red cell mass and intravascular volume, the anesthesiologist attends to plasma oncotic pressure, electrolyte composition, serum glucose, clotting factor levels, platelets, etc. In many cases, maintaining intravascular volume and management of coagulopathy are intertwined, so these topics are treated together in this section. Figure 57-14 gives a high-level overview of a strategy for volume and transfusion management as a function of blood loss during transplantation.
High-level overview of volume and blood component replacement strategy for liver transplantation requiring massive transfusion. All fluids are warmed (see text). The thickness of the bars represents the approximate relative proportions of each fluid in the replacement infusions.
Plasma oncotic pressure is a function of the osmotically active species in the plasma (proteins and, to a very small degree, cells). Albumin and other plasma proteins, mostly synthesized by the liver, provide plasma oncotic pressure. During liver transplantation, colloids should be used to replace intravascular volume lost through hemorrhage, with the addition of formed blood components as needed to meet specific needs (eg, red blood cells for oxygen-carrying capacity).
One attractive approach to manage plasma oncotic pressure is to use fresh frozen plasma as a volume expander. This is particularly appropriate when the patient is coagulopathic due to hepatic synthetic failure, as fresh frozen plasma contains most of the proteins found in normal plasma. However, fresh frozen plasma also contains most of the citrate added to blood at the time of collection to prevent clotting. Thus rapid infusions (ie, >1 mL/kg/min) of fresh frozen plasma can chelate Ca+2, causing acute hypocalcemia with consequent vasodilation and hypotension at inopportune times.210 Figure 57-15 illustrates the consequences of rapid infusion of citrated blood products during liver transplantation.
Acute systemic hypotension due to rapid transfusion of a citrate-containing 50:50 mixture of fresh frozen plasma and packed red blood cells. Boluses of 100 mL (vertical bars) were infused at 500 mL/min using a rapid infusion system (Haemonetics, Braintree, MA, USA). The hypotensive response is superimposed on acute hypotension due to sudden hemorrhage that was to be treated by the fluid challenge. The infusion-exacerbated hypotension was treated by infusion of norepinephrine.
Fresh frozen plasma is a poor choice for volume expansion when the patient is not coagulopathic, as each unit of fresh frozen plasma carries with it the potential to trigger transfusion-related lung injury. Also, liver transplantation entails the construction of multiple low pressure, low flow anastomoses. Some practitioners believe that maintaining a slightly hypocoagulable state lessens the possibility of unwanted hepatic or portal venous thromboses. Thus, in the absence of significant coagulopathy, 5% albumin solution is probably a better choice for volume expansion. Synthetic colloids comprised of long-chain polysaccharides are available but are not popular because they can interfere with clot formation. Additionally, anesthesiologists must be cognizant of not volume overloading the patient, as excessive volume postreperfusion can add to liver congestion and swelling within the capsule, likely amplifying ischemic reperfusion injury.
Transfusion and Other Therapies to Maintain Coagulation Capacity
Coagulation is monitored using standard laboratory tests, augmented in some cases by point-of-care functional tests, and ultimately by clinical correlation of the laboratory test results with direct observation of the surgical field. The basic tests to monitor coagulation status are the prothrombin time, the activated partial thromboplastin time, and the platelet count. D-dimer and fibrinogen levels provide information about the presence of clot lysis, and the latter, the ability to form clot. However, these tests provide a far-from-complete picture of a patient's coagulation performance. Many centers use thromboelastography to obtain near-patient diagnostics of clot initiation, formation, and lysis.211,212 Thromboelastography is a mechanical test described in detail in Chapter 15. The test is sensitive to extraneous influences (including environmental disruption), and it is not well standardized. However, thromboelastography is useful in liver transplantation because, unlike other tests, it gives information about clot lysis.212
Coagulation performance during liver transplantation in the patient with end-stage liver disease tends to follow a predictable course.213 Patients with hepatic synthetic failure start out hypocoagulable, and with appropriate transfusion of fresh frozen plasma as part of the volume replacement strategy, coagulation status tends not to get significantly worse during the preanhepatic and anhepatic phases of the operation. After reperfusion of the graft, a clot lysis syndrome develops, and some patients become hypercoagulable at the same time (ie, disseminated intravascular coagulation develops).213 Even patients who present with normal hepatic synthetic function and no coagulopathy can develop a dilutional coagulopathy as well as the reperfusion clot lysis syndrome. A worsening coagulopathy postreperfusion despite adequate replacement of coagulation factors, platelets, and fibrinogen may herald a nonviable graft and the immediate need for a replacement organ.
Volume replacement therapy should attend to preserving or moving clotting potential toward a normal state as part of the effort to reduce blood loss. However, transfusion therapy and coagulation management are no substitute for good surgical technique. Successful transplantation requires attention to both. Typically, fresh frozen plasma is used to replace clotting factors. Liver transplant patients are frequently in a state of low-grade disseminated intravascular coagulation and thus consume fibrinogen and clotting factors even in the absence of massive hemorrhage. Fresh frozen plasma usually supplies sufficient fibrinogen, but cryoprecipitate may be helpful if the fibrinogen levels become unacceptably low.
Platelets are rarely needed prior to graft reperfusion, even in patien ts with low platelet counts (eg, 50000/ml). Furthermore, many patients manifest splenic sequestration of platelets and have no response to transfusion, both exposing the patient to unnecessary risk of transfusion and wasting a valuable blood component.
Antifibrinolytic drugs and other procoagulants can be used prophylactically in the preanhepatic phase to minimize bleeding. These drugs are discussed in detail in the section on the postanhepatic phase of liver transplantation.
The anhepatic phase of liver transplantation is frequently described as the time from explantation of the diseased liver to the reperfusion of the new liver. However, the anhepatic phase of the operation actually begins once the native liver is sufficiently compromised to have no further function. During liver transplantation, this is usually when the hepatic artery is clamped, which can precede portal and hepatic vein clamping by many minutes. During this time, especially in the setting of portal hypertension, the diseased liver receives no oxygenated inflow and begins to die, with a consequent effect on the patient's acid–base status. However, once the liver is completely excluded from the circulation, this problem ends. After this, the anhepatic phase is often a relatively quiet period in the case. Bleeding should be under control, and the surgeons are engaged in the delicate task of constructing the venous anastomoses. During the construction of the venous anastomoses, the liver is kept on ice (Fig. 57-16, left panel).
Liver transplantation surgery. Left panel illustrates construction of the initial anastomosis, with the liver kept on ice to minimize warm ischemic time. The right panel illustrates the liver being placed into the body cavity. At this point, immunosuppressants should be given. The anastomoses are partially tied, after which the liver is flushed and the anastomoses opened.
Any drugs dependent on hepatic metabolism begin to accumulate once the portal vein and hepatic artery are clamped. Thus infusions of drugs should be titrated to effect. Of course, drugs not dependent on hepatic metabolism (eg, inhaled agents, remifentanyl) are unaffected by this transition.
The anesthesia team should attend to immunosuppression during the early anhepatic phase. The patient will have received an induction dose of an immunosuppressant, such as mycophenylate mofetil, prior to coming to the operating room. Obviously, this should be confirmed prior to starting the case. Typically, a modest dose of steroid (eg, 100-500 mg methylprednisolone) is given in the operating room as the new liver is initially placed into the recipient (before opening of the anastomoses; Fig. 57-16, right panel). For patients with hepatitis C, alternative regimens may be used such as an infusion of rabbit antithymocyte globulin begun at the completion of the hepatectomy and 10 mg methylprednisolone when the liver is placed into the recipient. In practice, the immunosuppression regimen is chosen by the surgeon, and the anesthesia team should ascertain the drugs, doses, routes, and timing prior to starting surgery.
Hemodynamics frequently stabilize during the anhepatic phase, particularly if a piggyback technique is used. This period of stability may derive from the fact that the rapid alterations in venous return during gross manipulations of the liver are superceded by more delicate tasks. If an en-bloc resection has been used with a complete vena cava cross-clamp, the circulation may still be stable if venovenous bypass is also used. However, if the vena cava is clamped without a venous return conduit, it is often necessary to support the circulation with a potent vasoconstrictor, preferably an agent with some inotropic activity such as norepinephrine.
Hypotension with diminished central venous pressure may occur even with significant preservation of venous return during the anhepatic phase. Lower extremity and splanchnic congestion lowers the effective circulating volume and central venous pressure above the diaphragm, with attendant hypotension. However, it is important not to overtreat this condition with aggressive volume replacement, as it can lead to hypervolemia at the time of graft reperfusion. Instead, it is preferable to support the circulation with a low dose of a vasopressor, anticipating the mobilization of blood from below the diaphragm once the liver is reperfused.
Finally, some poorly understood physiologic feature of the anhepatic phase might directly affect left ventricular performance, contributing to hypotension. Echocardiographic studies of ventricular function demonstrate, for example, that left ventricular shortening fraction diminishes during the anhepatic phase, relative to the preanhepatic phase, then returns to normal214 after reperfusion. Although this was a small study, a more recent and larger study revealed a decrease in right ventricular ejection fraction from baseline during the anhepatic phase of liver transplantation in 20 patients, but with a gradual return to baseline 30 minutes postreperfusion.215
Serum electrolytes and acid–base balance are subject to wide and rapid swings during liver transplantation, especially if the case is bloody and extensive transfusion is required. These are most severe during the anhepatic phase, when even the small residual function of the native liver has been removed. However, even during the preanhepatic phase, the patient's acid–base, electrolyte, red cell mass, and coagulation statuses are likely to change rapidly due to brisk hemorrhage, various cross-clamping maneuvers, and compression of abdominal vasculature, as well as potentially massive transfusion. Thus frequent monitoring of arterial blood gases, serum electrolytes, red cell mass, and laboratory measures of clotting is mandatory from the beginning of surgery.216 During the active portion of the case, we interpret and respond to laboratory results as they are returned, allow for equilibration, and then recheck the laboratory values.
Metabolic acidosis may develop in the anhepatic phase due to portal cross-clamping, as well as partial or complete inferior vena cava (IVC) cross-clamping, creating imperfect lower extremity and splanchnic venous return. There is also loss of any residual ability to clear organic acids. At the time of reperfusion, the donor liver and underperfused tissues from the patient release acid loads. Ongoing transfusion of low-pH, citrated blood products also contributes to the acidemia.
Various maneuvers can mitigate acidemia during liver transplantation. Hyperventilation can be used to establish an acute respiratory compensation for metabolic acidosis. Sodium bicarbonate can be administered to correct acidemia, although the need to minimize sodium loads, particularly in patients who are hyponatremic, may limit the usefulness of bicarbonate. Sodium bicarbonate must not be coadministered in the same infusion as Ca+2 solutions, as the combination will precipitate. Tromethamine (THAM) can also be used to control acidemia.
Acidemia at the time of reperfusion should be corrected by hyperventilation and administration of bicarbonate. This minimizes the hemodynamic and cardiac effect of the acid load and metabolic waste products released by reperfusion of the allograft. However, overzealous correction of metabolic acidemia is undesirable, particularly in the setting of massive transfusion. This is because the new liver clears citrate and organic acids, resulting in a rebound metabolic alkalosis; this may be seen developing even during the late neohepatic phase. Thus a reasonable strategy is to use modest hyperventilation to keep the pH near normal, with the use of bicarbonate of THAM only as needed to keep the pH above approximately 7.35 after respiratory maneuvers have been fully implemented.
Patients presenting for liver transplantation may be either hyper- or hypokalemic, depending on whether potassium-sparing or potassium-wasting diuretics have been used to control ascites. Potassium balance is also influenced by the underlying renal function or compromise. Serum potassium is subject to strong influences favoring hyperkalemia during the anhepatic phase. This may lead to large swings in serum potassium, especially if renal function is compromised or if large volumes of blood products are required. Additionally, splanchnic ischemia during the case tends to elevate serum potassium. Metabolic acidosis tends to worsen hyperkalemia. Finally, the preservative solution in the donor liver is rich in potassium.
Dangerous levels of serum potassium should be treated to avoid myocardial irritability and cardiac dysrhythmias. In patients with preserved renal function, this is accomplished initially by controlling the pH, by the use of potassium-wasting diuretics if the volume status permits, and by avoidance of potassium-containing solutions. Bicarbonate to increase pH tends to drive potassium intracellularly, whereas Ca+2 administration to replete deficits stabilizes irritable myocardium. Rapid or persistent increases in serum potassium are treated with glucose and insulin to drive potassium intracellularly by glucose cotransport. The combination of insulin and glucose predictably reduces serum potassium even during the anhepatic phase.217
Inhaled β2-agonists, including albuterol and levalbuterol, may be used to lower serum potassium by driving serum potassium intracellularly, and recently salbutamol was used to treat hyperkalemia resistant to other treatments during the anhepatic stage of a patient undergoing liver transplantation.218
If renal function is severely compromised, renal replacement therapy, either by continuous venovenous hemofiltration or by conventional dialysis, may be warranted to control serum potassium, sodium, and pH. Intraoperative renal replacement therapy can strain the resources of hospital dialysis services, so early consultation with a dialysis nephrologist is important once it appears that such therapy is needed.
Calcium is required for smooth muscle and myocardial contractility. Thus hypocalcemia during liver transplantation depresses inotropy (manifested as depressed cardiac index, reduced cardiac stroke index, and left ventricular work index219) and reduces vascular tone. In this setting, vasopressor activity is also compromised. Hypocalcemia may also interfere with clotting, as calcium is a required cofactor for many clotting factors. Thus it is important to monitor and correct abnormal serum calcium. During liver transplantation, serum ionized calcium tends to decrease due to vigorous administration of citrated blood products. As a rule of thumb, administration of 6 units or more of acid citrate dextrose (ACD)-preserved packed red blood cells requires supplementation of calcium.
Calcium can be repleted with calcium gluconate or calcium chloride. The positive inotropic effects of calcium administration are quick in onset but short lived. Again, overaggressive correction of ionized calcium intraoperatively precipitates a rebound hypercalcemia during the postoperative period as the liver metabolizes the citrate chelator. Thus the goal for intraoperative calcium management is to achieve the lowest ionized calcium concentration consistent with good cardiac performance and coagulation, typically 0.9 to 1.0 mmol/L. Postoperative hypercalcemia is treated with hydration and furosemide diuresis.
Serum sodium is frequently low in liver transplant patients due to the formation of ascites, the effects of renal failure, the use of diuretics, or combinations of all of these. It is important to avoid accidental aggressive overcorrection of hyponatremia, as too-rapid correction of hyponatremia can cause central pontine myelinolysis, an irreversible neurologic injury. Unfortunately, many useful replacement fluids contain sodium. For example, fresh frozen plasma, packed red blood cells, and 5% albumin all contain sodium at normal physiologic concentrations. To maintain intravascular volume while avoiding too-rapid correction of hyponatremia, a strategy of administering colloids along with 5% dextrose while inducing a diuresis of sodium and water with furosemide may be used. In patients with renal failure, dialysis or continuous venovenous hemofiltration against a low sodium bath can be used to remove the volume without increasing sodium too abruptly.
Glucose management during liver transplantation is usually fairly straightforward despite the liver's central role in glucose handling. The liver is the major site for gluconeogenesis, and this function is preserved to some degree in even the most severe cases of end-stage liver disease. Thus complete hepatectomy creates the possibility of intraoperative hypoglycemia during liver transplantation. This is particularly true if the anhepatic phase of the operation is prolonged.
Serum glucose during liver transplantation is also influenced by exogenous sources, including carrier infusions and the release of glucose from graft preservation solution. Thus intraoperative hypoglycemia is rarely a problem in practice, especially if some IV fluids (eg, drug-carrier infusions) contain 5% dextrose. Additionally, the use of methylprednisolone during the anhepatic stage often results in serum hyperglycemia. The liver is at once the major site of gluconeogenesis and insulin-mediated glucose uptake. Therefore, the anhepatic phase may be marked by hyper- or hypoglycemia, both of which should be avoided and treated. As in the case of other major surgeries, tight glycemic control is probably beneficial, and blood glucose should be measured frequently. Target blood glucose should be 150 to 200 mg/dL.
Renal Protection during Transplantation
Renal failure requiring renal replacement therapy (eg, dialysis or continuous venovenous hemofiltration) is common after liver transplantation, with rates of approximately 5% to 10%.220,221 Peritransplantation acute renal failure requiring renal replacement therapy is associated with higher mortality as compared with liver transplant patients not requiring dialysis.222 It is not clear whether there is a cause-and-effect relationship between acute renal failure and death after liver transplantation, but renal failure is an unwelcome outcome.
Protection of renal function probably depends, among other things, on maintaining renal perfusion during the liver transplant procedure. Renal perfusion is subject to many insults during transplantation, including hypovolemia and increased resistance to venous outflow due to vena cava compression or cross-clamping. Maintaining an adequate circulating volume facilitates preserving renal function. However, this is at times difficult, and a selective medical therapy to protect the kidney in the peritransplant period would be useful.
Attention has focused on selective splanchnic vasodilators, such as low-dose dopamine or, more recently, fenoldopam, as protective agents during major surgery. Fenoldopam has recently been studied in liver transplant patients randomized to 1 of 3 groups (fenoldopam, low-dose dopamine, or placebo) prior to surgery. Neither fenoldopam nor dopamine was superior to placebo for preservation of any measure of renal function (urine output, serum creatinine, creatinine clearance, use of diuretics, or use of pressors) immediately after surgery. On the third and fourth day after transplantation, creatinine clearance was reduced in patients receiving placebo or low-dose dopamine, whereas it was preserved in patients receiving fenoldopam.223 The authors concluded that fenoldopam might counteract the renal arterial constrictive effects of cyclosporine.223
In this small, unblinded study, neither low-dose dopamine nor fenoldopam demonstrated obvious renal protective effects in the early perioperative period.223 Similarly, another small study by the same researchers demonstrated that fenoldopam led to better creatinine and blood urea nitrogen values at day 3 after liver transplant, but the functional effect of this result on the incidence of acute renal failure was not demonstrated.224 Slightly lower (but not statistically significant) BUN and creatinine levels were observed in OLT patients who received continuous infusions of fenoldopam, although increased splanchnic perfusion was seen in the study group as compared to placebo.225
More conventional approaches to renal protection include maintaining adequate circulating volume and perfusion pressure. Urine output is commonly used to judge real function in the operating room, and "target" urine flows of 1 mL/kg−1/h−1 are frequently sought. When the urine output falls below this target and the circulating volume and perfusion pressure are judged to be adequate, furosemide or mannitol are sometimes used to increase the urine output.
Preparation for reperfusion involves optimizing volume status and hemodynamic performance and preparing the operating room for a potentially chaotic reperfusion period. Potential distractions related to the anesthetic (eg, the need to redose drugs, change syringes, carrier infusions) should be addressed in advance. Laboratory studies should have been sent in time for the results to arrive 5 minutes prior to reperfusion. Constant communication between surgeon and anesthesiologist must be the rule rather than the exception, with each team informing the other well in advance of any major intervention or treatment.
Abnormalities in pH, serum potassium, or calcium should be corrected prior to reperfusion. Sodium bicarbonate, glucose and insulin, and calcium chloride should be prepared and ready for use. Some practitioners use combinations of these agents prophylactically prior to reperfusion.174
Vasopressors may be required at the time of reperfusion and should be available for immediate infusion. Among the commonly used vasopressors, there is no compelling or logical best choice, so institutional preferences prevail. Epinephrine, dopamine, and norepinephrine are all acceptable choices, although some practitioners may consider the use of phenylephrine or vasopressin. Vasopressors are also commonly used prophylactically at reperfusion, frequently as a small bolus such as epinephrine, 10 mcg.
Just prior to reperfusion, the new liver is flushed antegrade from the portal vein to the hepatic vein to wash out the preservative solution (which contains large doses of potassium and heparin). Flushing is accomplished with crystalloid, colloid (such as 5% albumin), or blood. A blood flush is performed by constructing the portal and hepatic venous anastomoses, after which the portal anastomosis is closed but the hepatic venous anastomosis is left open. The portal vein cross-clamp is removed, and blood is allowed to flush the liver and run out into the field where it is recovered by suction. Concomitant with this controlled hemorrhage, fluid should be rapidly transfused to maintain euvolemia.
Graft reperfusion has its most obvious effect on cardiac performance, and the anesthesia team should be prepared for pulmonary hypertension and acute right heart dysfunction. Liver transplantation in and of itself does not appear to cause right heart dysfunction.226 However, release of preservative solution, air, clot, and debris (as well as acidemic blood from the reperfused splanchnic circulation) into the pulmonary vasculature can cause sudden and severe pulmonary hypertension, elevated right heart pressures, and right ventricular failure (as mentioned previously). Systemic pressure (and coronary perfusion pressure) decreases due to inadequate left ventricular preload. The elevated right heart pressures may open an occult patent foramen ovale, with risk of paradoxical embolism.227 The central venous pressure frequently increases due to right heart and pulmonary congestion. Ordinarily this is desirable to a point to ensure right ventricular preload. However, in the immediate reperfusion period, elevated central venous pressure compromises graft perfusion and may contribute to early graft failure. Thus it becomes important to preserve or reestablish efficient right ventricular pumping and a low-resistance pulmonary circulation.
Pulmonary hypertension, right ventricular overload, and the resulting central venous congestion reduce the hepatic perfusion pressure, which is simply the difference between the portal and central venous pressure when only the venous anastomoses have been constructed. The portal venous pressure is not typically measured, so the hepatic perfusion pressure is not directly accessible. However, inadequate perfusion pressure can be deduced if the liver becomes engorged. This alarming appearance somewhat mimics that of hyperacute rejection. To avoid confusion, the hepatic perfusion pressure is ideally controlled by keeping the central venous pressure low at the time of reperfusion. This is not possible in the event of acute pulmonary hypertension unless inotropes and selective pulmonary vasodilators are used. Typically, the heart is supported with an inotrope, and the lungs are hyperventilated to raise the pH and induce vasodilation. Inhaled nitric oxide and/or a low-dose infusion of nitroglycerine may be useful at this point.
As stated previously, good communication between the surgical and anesthesia teams is also important to minimize the effect of reperfusion. Reperfusion should not occur until the anesthesia team has optimized the patient's physiology, and this should be accomplished in time to allow prompt reperfusion. In the event of severe problems at reperfusion, the surgeon can partially reclamp the portal vein to lessen the effect of the effluent from the new liver.
The combined effect of multiple insults on cardiac and pulmonary performance at reperfusion may lead to cardiovascular collapse. Advanced cardiac life support protocols directed at the presumptive major problem should be promptly initiated while the underlying insults are identified and corrected. In the event of cardiopulmonary arrest not responsive to pharmacologic support (or due to reversible causes such as pulmonary embolus), mechanical circulation and oxygenation can be provided by percutaneous femoral venoarterial cardiopulmonary bypass with acceptable survival.228
In the early postreperfusion period, severe volume overload from overaggressive fluid replacement during the period of poor venous return may compromise hepatic perfusion even with adequate cardiac performance. If the problem is slight, then it may be possible to administer furosemide to induce diuresis, or to wait for the blood volume to fall due to bleeding. In the case of severe volume overload and poor graft perfusion, phlebotomy from a large port on a central venous catheter is a logical option.
The neohepatic phase begins with the initial reperfusion of the liver. Frequently, the next step is to construct the hepatic arterial anastomosis. The liver ultimately needs a high-pressure, oxygenated perfusion source, and the surgical team will turn its attention to this as soon as the immediate effects of venous reperfusion have passed. Opening of the hepatic arterial anastomosis usually has little or no hemodynamic effect.
An acute clot lysis syndrome frequently develops in the early neohepatic phase. This manifests clinically as diffuse bleeding from previously coagulated sites in the surgical field, new oozing from previously quiescent vascular catheter insertion sites, and ongoing transfusion requirements. Laboratory analysis shows a further elevation of the prothrombin time relative to baseline, and a sometimes profound elevation of the partial thromboplastin time (eg, to >150 seconds), as if the patient had received a large dose of heparin. Thromboelastography demonstrates poor clot initiation and rapid dissolution of clot. Figure 57-17 shows sequential thromboelastograms obtained from a single patient during the course of transplantation.229
Sequential thromboelastograms from a single patient having liver transplantation. Marked fibrinolysis developed during the anhepatic phase. Note the development of severe coagulopathy in the "5 min after reperfusion" sample consisting of poor clotting and early fibrinolysis more severe than during the anhepatic phase. ε-Aminocaproic acid (EACA) was given, which arrested the clot lysis. [From Kang Y, Lewis JH, Navalgund A, et al. Epsilon-aminocaproic acid for treatment of fibrinolysis during liver transplantation. Anesthesiology. 1987;66:766-773. Reproduced with permission.]
The cause of this clot lysis syndrome is not completely understood, which limits the ability to tailor treatment to specific causes. Multiple mechanisms have been proposed and may operate simultaneously, including release of heparin (from preservative solution) during reperfusion or release of endogenous activators of tissue plasminogen activator and/or endogenous heparinoids from the graft.213
A central goal in this phase of the transplant is to stop clot lysis. A key first step toward achieving this goal is to provide adequate levels of platelets and factors to support clotting in the face of ongoing consumption, with the expectation that a functioning graft will begin to provide appropriate clot lysis inhibitors and clotting factors on its own. One could consider administering a small (20-50 mg) dose of protamine213 in the event that clot formation is suspected to be deficient, especially if the liver flush prior to reperfusion was not optimal and the patient might have received heparin from the graft.
Fibrinolytic activity is frequently increased during liver transplantation, reflecting reduced hepatic clearance of tissue plasminogen activating factor (tPA) during the anhepatic phase and release of tissue plasminogen activator from the epithelial cells of the reperfused graft.230-232 In addition, plasminogen activator inhibitor-1 activity is reduced231 during transplantation. Fibrinolysis probably contributes to the unwanted clot lysis syndrome after reperfusion.
Strategies for reducing fibrinolysis and transfusion requirements include the use of antifibrinolytic agents such as aprotinin,233 tranexamic acid, and ϵ-aminocaproic acid.234 However, the use of antifibrinolytics in liver transplantation is somewhat controversial because of concerns about unwanted thrombosis.
ϵ-Aminocaproic acid and tranexamic acid are both synthetic analogues of the amino acid lysine. ϵ-Aminocaproic acid and tranexamic acid competitively inhibit the binding of plasminogen (a lysine protease) to lysine residues on fibrin, thus inhibiting fibrinolysis. Both drugs also prevent the conversion of plasminogen to plasmin, again by virtue of acting as competitive antagonists of a lysine protease. ϵ-Aminocaproic acid is used during liver transplantation as a loading dose (typically a 1- to 2-g bolus over 1-10 minutes), followed by infusions of 1 to 2 g per hour. It reliably arrests fibrinolysis as assessed by thromboelastography (see Fig. 57-17).229 The drug is eliminated by the kidney, with 65% of a dose appearing unchanged in the urine. The elimination half-life of ϵ-Aminocaproic acid is about 2 hours.
Tranexamic acid is 6 to 10 times more potent than ϵ-aminocaproic acid and has a longer elimination half-life. It is also renally eliminated, with 95% of the drug appearing unchanged in the urine. It is the least well studied of the 3 major antifibrinolytics.234 Dose rates between 2 mg/kg/h and 40 mg/kg/h have been reported in small studies with either ϵ-aminocaproic acid or aprotinin as comparisons.
Aprotinin is a protease inhibitor isolated from the lungs of swine or cows. It inhibits a variety of human proteases, including plasmin, trypsin, kallikrein, chymotrypsin, activated protein C, and thrombin. These proteases are important in fibrinolysis and coagulation, as well as complement activation and inflammation. Each uses a serine residue for catalysis at the active site, and aprotinin acts by forming a specific aprotinin-active site serine complex with each protease. Aprotinin has an approximately 1-hour redistribution half-life after a bolus, and a 7- to 10-hour elimination half-life. It is given as a bolus, followed by an infusion. A typical bolus is 2 million kallikrein inactivator units (KIU), followed by an infusion of 500000 to 1 million KIU per hour.234
The use of antifibrinolytic drugs may have deleterious effects. There is anecdotal evidence that antifibrinolytic therapy may increase the incidence of thrombotic events during liver transplantation, including fatal pulmonary thromboembolism.235 Additionally, as mentioned previously, aprotinin is no more effective than the lysine analogues, but it is associated with excess morbidity and mortality relative to these other drugs when used in cardiac surgery.132 This suggests that during liver transplantation, antifibrinolytic agents should be reserved for patients who have a relatively high risk of severe hemorrhage,236 and aprotinin might best be avoided altogether. However, the clotting cascade in liver transplant patients is quite deranged, and cardiopulmonary thromboembolism, even in the absence of exogenous procoagulants, may be more common than is usually appreciated.169
Thus antifibrinolytics might be considered for rescue if bleeding becomes a problem in the postanhepatic phase,234 as judged by the disappearance of clot from the surgical field, or as a prophylactic measure in patients expected to have severe bleeding problems (eg, multiorgan transplant in a patient with renal failure). Balanced against this cautious approach are some practices that advocate the use of antifibrinolytics prophylactically, in advance of any pathologic bleeding.234 However, in addition to enhanced fibrinolysis, many other factors contribute to blood loss late in orthotopic liver transplantation, including coagulopathy due to synthetic failure, thrombocytopenia, platelet dysfunction, dysfibrinogenemia, dilutional coagulopathy, hypothermia, and bleeding from surgical technical problems. As postreperfusion fibrinolysis is largely a diagnosis of exclusion, all of these other factors must be ruled out before administration of potent procoagulants.
Procoagulant agents, such as purified specific clotting factors, can also be used, either prophylactically or as an attempted therapy for excessive bleeding. For example, in 1 study, recombinant factor 7 corrected coagulopathy when given as single bolus prior to transplant, but there was no effect on transfusion requirements.237 In a different study, recombinant factor VII was used in a small number of patients without clear evidence of clinical harm or benefit.238 Recombinant factor VII has the theoretical advantage of being selective only for areas that are bleeding—that is, those areas that have sustained tissue damage. This is because recombinant factor VII requires tissue factor for full activity, and tissue factor is only available in areas in which the subendothelial layers are exposed. Of course, the vascular anastomoses in the new graft would have exposed tissue factor, at least theoretically increasing the risk of thromboses of the anastomoses. Furthermore, recombinant factor VII is apparently not as selective as initially hoped. A recent report documents many episodes of arterial and venous thromboses distant from the hoped-for site of action (eg, cerebrovascular accident, myocardial infarction, pulmonary embolism, and deep venous thrombosis) when recombinant factor VII was used in various "off-label" applications.134
If the fibrinolysis reaction cannot be controlled, the patient will have an ongoing transfusion requirement. The anesthesia team should quantify the hourly transfusion needs in the neohepatic phase and make preparations to meet this during the end of the case, the move from OR table to ICU bed, transport to the ICU, and in the intensive care unit itself.
L-arginine, an essential component of the L-arginine/nitric oxide (NO) synthase pathway, is diminished in patients postreperfusion due to liberation of arginase from the donor graft, and may lead to hemodynamic changes in the reperfusion stage.239,240 With a decrease in L-arginine and an increase in arginase, there is a concomitant increase in L-ornithine levels.241 Blockade of the L-arginine/NO synthase pathway is implicated in hepatic apoptosis and liver transplant preservation injury, and blockade of arginase or administration of L-arginine protects against hepatic warm ischemia reperfusion injury and has demonstrated improvements in cardiac output and liver blood flow, while reducing pulmonary vascular resistance.242-244
During the neohepatic phase, the anesthesia team prepares for the postoperative disposition of the patient. Most liver transplant recipients are discharged from the operating room to an ICU bed. The anesthesia team seeks to manage this transition as smoothly as possible. The patient should be transported on an ICU bed, with all necessary infusions running and the patient fully monitored. Comprehensive communication between the operating room and the ICU team is essential to facilitate the smooth transition of care.
The anesthesia team also prepares for the early postoperative care of the patient. One large, multicenter study looking at early extubation in 391 patients demonstrated that early extubation is safe under specific conditions, with only 7.7% having adverse outcomes secondary to either pulmonary issues or need for additional surgery.245 Suitability for extubation is influenced by many factors. Among these are the intraoperative transfusion requirement (as a warning of possible airway edema); the presence or development of pulmonary, cardiac, and renal compromise; and the presence and severity of encephalopathy.246 Fluid administration and possible volume overload are some of the many factors that can interfere with the patient's readiness for extubation at end of case. Controlled fluid administration supported by adjuvant vasopressors led to reduced rates of reintubation in 1 study.247 Intraabdominal hypertension is common after orthotopic liver transplantation and limits the potential for extubation. Intraabdominal hypertension (pressure >25 mm Hg in the bladder) developed in 32% of patients248 in 1 series. It is expected that these patients are less likely to meet criteria for early extubation and are harder to wean from the ventilator.
If in doubt, it is always acceptable and prudent to leave the patient intubated for pulmonary support in the early postoperative period. Postoperative mechanical ventilation allows the team to focus on ongoing transfusion requirements, to exclude the possibility of intra-abdominal hypertension, and to gradually prepare the patient for extubation. Despite the concern about elevated intra-abdominal pressures on graft viability, postoperative application of positive end-expiratory pressure (PEEP) does not have a major effect on graft function, although in half of the patients the cardiac output declined with PEEP.249 This is encouraging because PEEP may be needed to counteract the pulmonary consequences of intra-abdominal hypertension.
Pain management is relatively straightforward after liver transplantation. Many patients remain intubated and receive potent opioids as sedatives. Epidural analgesia is traditionally avoided due to the profound coagulopathy manifested by end-stage liver disease patients, both preoperatively and in the postoperative period, although some centers report successful use of thoracic epidural anesthesia for patients who meet strict criteria.250 Furthermore, orthotopic liver transplant patients have less postoperative pain than hepatectomy patients.251 The reason for this difference in pain threshold is unclear, although some have implicated the steroids used for immunosuppression. Steroids have analgesic properties in other patient populations.252